Start here. An illicit financial flow is money in illicit motion. Before the types, the sectors, the drivers and the methods that come later, two ideas have to be solid: what makes a flow illicit, and what it means to call it a flow.
A flow is illicit when it is illegally earned, illegally moved, or illegally used. The illegality can sit at any one of those three points. It does not have to fail all three. Tap each to see it.
Now the second idea. IFFs are flows: money in motion, not money sitting still. And they move in two directions that are easy to confuse. Tap each.
This is the form most people picture, and the form the standard definition stresses. Value leaves one country for another in a way the law forbids.
Example. A mining company exports a consignment of ore but declares it on the customs paperwork at half its true value. The foreign buyer pays the real, higher price into a bank account the company controls offshore. The ore has gone; half of its value never came home, was never recorded, and will never be taxed in the country that owned the resource.
Not every illicit flow crosses a border. Illicit value usually moves internally first, and some of it never leaves the country at all.
Example. A government official approves a public road contract at three times a fair price. The contractor builds a cheaper road and quietly pays the surplus to a company the official secretly controls, registered in the same country. No border has been crossed, yet public money has moved illicitly from the national budget into a private pocket.
Why both matter. The internal flow usually feeds the cross-border one. Sooner or later, the official’s company will try to move that money offshore. Internal and cross-border are two stages of a single pipeline, so seeing the whole problem means watching both.
Treat these as orders of magnitude. Every IFF figure is an estimate built from gaps in data, not an audited count.
We have seen what makes a flow illicit and that it runs both across and within borders. The next step is to tell the types apart, because each type is generated, detected and stopped in a different way. There are five.
Tap each type.
Here the harm lies in escaping tax. This type covers two things the law treats differently but that cost a country exactly the same revenue.
The first is outright tax evasion: illegally hiding income, profit or assets so they are never declared. The second is aggressive tax avoidance and abusive transfer pricing, where a multinational shifts taxable profit to a low-tax place through arrangements that may be lawful on their face yet defeat the purpose of the law. Evasion is a crime; abusive avoidance is a grey zone. Both belong here, because the fiscal loss is identical.
How it moves. Profit shifting, mispriced transactions between related companies, hidden offshore income, intra-group loans and royalties that strip a local company’s taxable profit, and the abuse of tax holidays and exemptions.
Here the illegality lies in the trade itself. The defining mechanism is trade misinvoicing: an exporter under-declares the value of a shipment, or an importer over-declares it, so value crosses the border hidden inside ordinary, correct-looking trade documents.
It also includes outright smuggling, where goods cross with no declaration at all; customs valuation and misclassification fraud, where goods are undervalued or booked under the wrong tariff code; and transit fraud, where goods declared as merely passing through a country are diverted and sold there instead.
How it moves. A fuel tanker enters a country declared as duty-free transit to a neighbour. The fuel is quietly offloaded and sold on the local market; the tanker is refilled with water and driven on to the exit border to be signed off as having left. The duty was never paid, and the trade documents look perfectly ordinary.
Here the money is the proceeds of crime, and it is dirty from the very start. Drug and human trafficking, smuggling, fraud, illegal mining, and wildlife and timber crime.
The aim of the flow is not to escape tax; it is to launder the money: to wash a criminal origin out of it by cycling it through trade, property and the banking system until it can be spent and invested in the open.
How it moves. Cash is placed into the financial system in small amounts, layered through many accounts and companies to break the trail, then integrated back as apparently clean investment, often in property or a legitimate-looking business.
Here the money is public money, taken by those who hold power. Bribery and kickbacks, inflated public contracts, ghost projects, and the embezzlement of state funds and the assets of state-owned enterprises.
By value, corruption is usually the smallest of these channels. But it does the deepest damage, because it strikes directly at the trust between a state and its citizens, and because it is the type most often protected by political influence.
How it moves. It very often begins as an internal flow, public money diverted into a domestic company, and only later becomes a cross-border flow, when the proceeds are moved offshore and hidden.
This type is different in kind, and naming it carefully matters. The first four types describe where the illegality comes from: a tax, a trade, a crime, an act of corruption. This fifth one describes the medium the money moves through.
Crypto-assets, mobile money, online payment rails, digital platforms and online gambling have given each of the other four a faster, cheaper and harder-to-trace way to travel. So the most accurate way to describe it is not as a fifth source of illicit money, but as a fifth highway, and one the other four increasingly all use.
How it moves. The proceeds of a crime are converted into a crypto-asset, split across dozens of digital wallets, and cashed out in several countries within hours, faster than any paper-based control can follow.
The five types tell you what kind of flow you are dealing with. The sectors tell you where to look for it. IFFs concentrate where value is high, where a fair price is hard to verify, and where goods or money are easy to move.
Tap a sector to see how IFFs occur there, the type involved, and an example.
We now know what IFFs are, the five types, and the sectors they run through. The hard question is why they continue. Each driver below carries an example, and two of them, the professional enablers and the secrecy jurisdictions, get a worked case. Tap each.
An IFF is, for the person behind it, a rational decision. The reward is large and immediate; the chance of being caught is low; and the penalty, when it does come, is often smaller than the gain it punishes.
Example. A single misinvoiced shipment can move millions of dollars of value across a border in an afternoon. If it is caught, which most are not, the fine may be a fraction of the sum involved. Until that balance shifts, through a real chance of detection and a penalty that genuinely bites, the flow keeps making sense to the people running it.
Many flows are possible because the law leaves the door open: no transfer pricing rules, or only thin ones; no general anti-avoidance rule; no requirement to disclose who really owns a company; exemptions drawn so wide they invite abuse.
Example. A country with no transfer pricing rules has no legal basis to challenge a multinational that books its profit in a low-tax jurisdiction. The conduct is not illegal, so nothing is owed, and the loss is permanent. Where the law is silent or ambiguous, the loss is identical to theft, but lawful.
This is the driver people most often miss. Large IFFs are not built by amateurs. They are designed, documented and operated by professionals: lawyers, accountants, auditors, bankers, company-formation agents and corporate service providers. They draft the structures, sign off the accounts, open the bank accounts and supply the nominee directors. The enabler is not a bystander to the flow. The enabler is the machinery of it.
In 2016 a leak of 11.5 million documents from a single Panamanian law firm, Mossack Fonseca, exposed how one firm had created more than 210,000 offshore companies for clients across the world. It had industrialised secrecy, selling shell companies almost the way a shop sells goods. The most telling detail: when the leak broke, the firm itself could not identify the real owners of more than 70 per cent of the active companies it ran in its busiest hub. The people who built the structures did not know, and did not need to know, who was behind them.
The lesson: you cannot curb IFFs while leaving the enabler industry unregulated. The structures do not assemble themselves.
Some places sell secrecy as a national product. They offer low or zero tax, no requirement to reveal who owns a company, fast and cheap incorporation, and strong legal protection for that opacity. They are where the trail is meant to go cold.
More than half of the 210,000 companies in the Panama Papers were incorporated in the British Virgin Islands, a Caribbean territory with a population of only a few tens of thousands of people. A jurisdiction that small has no domestic need for hundreds of thousands of companies. It hosts them because corporate secrecy is, in effect, its export industry. The shell company is registered there precisely so that an investigator in the country losing the money hits a legal wall.
Secrecy jurisdictions are why an IFF that begins as a visible transaction can, within two or three steps, become untraceable.
Inside one government, customs, the revenue authority, the company registry, the central bank and the line ministries each hold one piece of the picture, and rarely combine them.
Example. Customs records a shipment declared at one value. The revenue authority holds the same company’s accounts, which imply a different value. The registry knows who the directors are. No single office sees all three, so the contradiction between them is never noticed, and the flow escapes in the gap between two mandates.
The methods that detect IFFs are well established. What is often missing is the capacity to run them, and the data to run them on.
Example. A revenue authority may have one or two trained transfer pricing specialists, facing the combined tax-advisory teams of every multinational operating in the country. Add trade, production and ownership data that is not connected, not timely or not trusted, and even a sound method has nothing to work with. Both staff and clean data are budget decisions, not technical ones.
Some flows are protected. Where the people benefiting from an IFF are powerful, rules that exist on paper are quietly not enforced.
Example. An audit that begins to point towards a politically connected company is not formally blocked; it simply slows, loses its staff, and never reaches a conclusion. This is the driver no detection method can reach. It is addressed only by transparency, by independent institutions, and by naming it honestly rather than treating IFFs as a purely technical problem.
An IFF is, in part, the distance between what really happened and what was recorded. Where activity is not recorded at all, there is no baseline, and value can move with nothing to compare it against. Several overlapping “hidden economies” create that blind space.
Example. A country cannot detect under-declared cattle exports if it never recorded how many cattle existed. With no baseline, the loss is not even a gap; it is an absence. This is why bringing activity into the record, registering traders, valuing care work, measuring production, is itself an anti-IFF measure.
We have the anatomy and the drivers. Now the practical question: how is an IFF found? The next two steps set out the methods, and each one is given the same three parts, the method, how to run it, and a real case. Tap each.
The method. Every cross-border trade is recorded twice: once by the exporting country, once by the importing country. The two records should broadly agree. A persistent gap signals misinvoicing. Where one side records the trade and the other records nothing, the signal is smuggling.
How to run it. Take bilateral trade data at the six-digit product level. Pair what country A reports exporting of a product to country B with what country B reports importing of the same product from A. Adjust for the standard valuation difference between an export value and an import value, roughly 10 to 12 per cent for freight and insurance, and for timing differences around year-end. Rank the residual gaps by size and by how consistently they recur, and refer the largest persistent gaps for audit.
The case: gold into the UAE. An analysis of UN trade data (SwissAid, 2024) found that about 2,569 tonnes of gold entered the United Arab Emirates from Africa over 2012 to 2022 with no matching export declaration from the source countries, worth roughly US$115 billion. The mirror gap was the whole story: the UAE recorded receiving the gold; the source countries recorded sending almost none of it. For smuggled gold the gap is near-total, because it never enters the source country’s export data at all. The method does not catch the courier. It sizes the leak and names the destination.
The method. Quantities can match while values are wrong. The price filter compares the declared unit price of each shipment, its value divided by its quantity, against a benchmark band of normal world prices for that exact product.
How to run it. For a chosen product, build a reference price from world trade data, an average or an interquartile band. Compute the unit price of every export and import transaction in that product. Flag any transaction priced outside the band: under-priced exports and over-priced imports are the ones that move value out. Weight the flagged transactions by volume to size the exposure, then pull the customs documents on the largest of them for a valuation audit.
The case: under-priced commodity exports. Global Financial Integrity’s value-gap studies have repeatedly found mineral and agricultural exports declared at unit prices well below the world reference for the grade. Run a price filter over a year of consignments and a cluster falls far below the band; the audit that follows typically shows the buyer is a related trader, and the distance between the declared price and the reference price is the profit shifted out. The filter does not prove intent. It tells the auditor which 20 entries out of 20,000 to open.
The method. Do not rely on trade documents alone, because the documents are exactly what a misinvoicer controls. Instead, compare what was physically produced, from mine output records, production licences and known capacity, against what was declared as exported and what actually reached the treasury. The difference is value that left between the pit and the books.
How to run it. Assemble production data from the line ministry, the regulator and the producer. Convert it to expected export volume and value at reference prices. Set that against declared exports and recorded fiscal receipts. Investigate the gap, and the specific points, the mine gate, the valuation step, the marketing channel, where the chain of custody breaks.
The case: the Marange diamond fields. In 2016 the President of Zimbabwe stated publicly that the country may have lost in the order of US$15 billion in diamond revenue from the Marange fields, with the treasury receiving under US$2 billion. The figure is contested and was later partly walked back. But the structure of the problem is the lesson: very large production, only a small fraction reaching the treasury, and a reconciliation gap that pointed straight at the valuation and marketing steps. Reconciliation works because it does not trust the export paperwork; it tests it against the rock that came out of the ground.
The method. Some IFFs are too diffuse to catch one transaction at a time. Macroeconomic methods estimate unrecorded outflows for the whole economy; reporting analysis reads the standardised data that multinationals now have to file.
How to run it. The balance-of-payments residual method, sometimes called net errors and omissions, compares recorded inflows and outflows of capital; a large, persistent, unexplained residual is a signal of unrecorded movement. Country-by-country reporting analysis takes the report each large multinational group now files, breaking down its revenue, profit, tax, staff and assets country by country, and looks for the mismatch: profit booked where there are few employees and little real activity.
The case: profit without people. Country-by-country data has repeatedly shown multinational groups booking enormous profit in small, low-tax jurisdictions that have almost no employees, set against large operations, with many staff, in higher-tax countries that report little profit. The mismatch between where the profit sits and where the people and activity are is the signal. This method turned profit shifting from something argued case by case into something visible in a single table.
The first four methods read data the state already holds. These next four test the price between related parties, the people hidden behind a structure, the movement of the money itself, and what is physically inside a truck. Tap each.
The method. When related companies in the same group trade with each other, the price between them is set, not discovered in a market. Comparability analysis tests that internal price against the price independent parties would have charged for the same thing. That benchmark is the arm’s length standard.
How to run it. Identify the controlled transaction and what the local company actually does, the functions it performs, the assets it uses and the risks it bears. Choose a transfer pricing method. Build comparables from genuinely independent transactions, or from published reference prices. For a mineral, start from the world metal price and subtract only arm’s-length treatment and refining charges. Compare the result to the price the company used. A price below the benchmark is profit shifted out, and the shortfall is an assessable adjustment.
The case: semi-processed metals. Many countries export platinum or copper as concentrate to a refinery that belongs to the same corporate group. The taxable question is the value of the metal contained in that concentrate. Benchmarking the inter-company price against published metal prices, less arm’s-length charges, lets the revenue authority test whether the full contained value was recognised at home. Where it was not, the difference is recovered as a transfer-pricing adjustment.
The method. An IFF structure hides the real owner behind layers of companies, trusts and nominee directors. Tracing works backwards through the layers to the natural person who ultimately controls and benefits.
How to run it. Pull the company registry record, then the record of each shareholder company in turn. Cross-match directors, registered addresses, phone numbers and signatories across the entities; the same details recurring across “unrelated” companies is the giveaway. Use leaked datasets and commercial corporate databases. A verified, public beneficial ownership register, where one exists, collapses this work from months into minutes.
The case: the Panama Papers structures. The 2016 leak showed why this method matters. Tens of thousands of companies, each looking separate, resolved, once ownership and signatories were mapped, into networks controlled by a far smaller number of real people, including public officials and their associates. Investigators since have shown the same pattern repeatedly: a list of unrelated-looking exporters or buyers, traced through the registry, turns out to be one network with one owner.
The method. Follow the money through the financial system. Banks and other institutions file suspicious transaction reports; a Financial Intelligence Unit analyses them for the patterns that betray laundering, above all layering, the rapid movement of money through many accounts to break the trail between origin and destination.
How to run it. The Financial Intelligence Unit receives suspicious transaction and large-cash reports and builds the network of accounts, counterparties and timing. Analysts look for structuring, amounts kept just below the reporting threshold; for funds moving faster than any real trade would require; and for counterparties with no economic relationship to the customer. A developed file is referred to the revenue authority or the police with the money trail already drawn.
The case: from courier to network. A single interception, a courier stopped at an airport with undeclared gold or cash, is only the visible end of a flow. Transaction analysis is what links that courier to the accounts that financed the consignment and would have received the proceeds, and so to the network behind it. The laundering investigations that followed the major gold-smuggling exposures of recent years were built on exactly this.
The method. Not every IFF is found in a spreadsheet. Some are found by looking, with cargo scanners, electronic cargo tracking, drones and satellite imagery, at what is physically moving and where.
How to run it. Cargo scanners let customs see inside a sealed truck or container without unloading it, exposing concealed compartments and cargo that does not match its declaration; paired with risk profiling, so the right trucks are scanned, they turn misdescription from a safe gamble into a likely detection. Electronic cargo tracking seals transit goods to fixed routes and flags any deviation. Drones and satellite imagery watch borders, rivers and mining sites no officer can physically patrol.
The case: drones on the river. A revenue authority placed surveillance drones, carrying night vision and high-resolution cameras, over a river border heavily used for smuggling. In one operation, with other agencies, it located and destroyed 21 boats and 20 paddles used to ferry contraband across the river, most of them found from the air. The lesson is general: a river is a border, and aerial surveillance turns an un-watchable stretch of it into an evidenced one.
These are the principal methods, not the whole list. IFFs are also uncovered by forensic tax audits, by whistleblowers and large data leaks, of which the Panama Papers is the best known, by open-source and network analysis, and by customs risk-profiling systems. The methods overlap, and a real investigation almost always combines several.
We can now name an IFF, place it in a sector, explain why it persists and detect it. The last question is what actually reduces it. No single measure works alone; the response has to match the drivers. Tap each.
Enact and keep current the rules the flows exploit: transfer pricing rules, a general anti-avoidance rule, and clear definitions of the terms disputes turn on. A flow is only illicit by reference to law. Where the law is silent, the loss is identical but lawful, and nothing can be recovered.
This is the single measure that strikes hardest at the enabler and the secrecy jurisdiction together. A register that records, verifies and publishes the real human owner of every company removes the anonymity the whole shell-company industry is built to sell. It turns months of ownership tracing into a search that takes minutes.
Two international standards now form the backbone of the response. Joining and actually using them, not merely signing up, is among the highest-value moves a country can make.
Automatic Exchange of Information (AEOI). Under the Common Reporting Standard, jurisdictions automatically and annually send each other information on financial accounts held by each other’s residents. A bank account hidden offshore is no longer invisible to the account holder’s home tax authority. More than 100 jurisdictions now exchange this information.
Country-by-Country Reporting (CbCR). Every large multinational group files a report breaking down, for each country, its revenue, profit, tax paid, employees and assets. Tax authorities receive it and can see at a glance where profit is booked away from real activity. AEOI attacks hidden offshore wealth; CbCR attacks profit shifting. Together they reach the secrecy and the mispricing that the domestic measures only chip at.
Hold lawyers, accountants, auditors, banks and company-formation agents accountable for the structures they build and the money they move. Real obligations to identify the true client, real penalties for failure, and professional consequences. The structures do not assemble themselves; the people who assemble them must carry the risk.
Require customs, the revenue authority, the company registry, the central bank, the financial intelligence unit and the line ministries to share data on a common basis. IFFs live in the gaps between mandates; closing those gaps is mostly a matter of law and political will, not technology.
Audit every tax holiday, exemption and Special Economic Zone against the revenue it costs and the behaviour it actually produces. Loosely policed incentives are an IFF channel in their own right, through round-tripping and origin fraud, and they are the cleanest channel to close, because the government controls them directly.
Beyond the automatic standards, use the exchange-of-information articles in tax treaties to make specific requests, run joint and simultaneous audits with other countries, and pursue mutual legal assistance and asset recovery. Example. Stolen public funds traced to property abroad can only be frozen and returned through asset-recovery cooperation with the country where the property sits. A secrecy jurisdiction can only be reached through cooperation.
Enforcement has costs as well as benefits, and the honest response owns that. A clear example: to stop fuel-transit fraud, a country may require importers to pay the duty up front and reclaim it only once the fuel is proven to have left, backed by electronic cargo tracking on fixed routes.
The measure cuts the fraud. It also raises costs for legitimate transit, and can push genuine trade onto a rival corridor in a neighbouring country. Revenue protection and corridor competitiveness pull against each other. A good response decides that trade-off deliberately and defends it openly, rather than pretending it does not exist.
This is the test of whether the picture has come together. Each scenario asks you to name a type, a sector or the right detection method. Tap your answer, then reveal it.
Mirror analysis and production reconciliation.
The partner comparison is a mirror-analysis signal. The production comparison is reconciliation. Together they size the leak and show it is smuggling, gold that never entered the export data at all, rather than mispricing. Transfer pricing tests a suspect price; here the problem is missing volume. Financial intelligence comes later, to trace the money once the leak is sized.
Country-by-country reporting analysis.
CbCR puts revenue, profit, tax, staff and assets side by side for every country. The mismatch here, enormous profit where there are three people, almost none where there are 4,000, is exactly the signal it is designed to expose. It is a tax-related flow: profit shifting. Transfer pricing analysis would then be used to quantify and assess it.
Corruption-related, and internal.
Public money has been stolen through an inflated contract, so the type is corruption. No border has been crossed yet, so the flow is internal. Watch what happens next: the official’s company will, in time, try to move the money offshore, and the internal flow becomes a cross-border one. That is the pipeline in action.
Beneficial ownership tracing.
Cross-matching directors, addresses and signatories across registry records is what turns a set of separate “exporters” into one network with one owner. It is exactly how the Panama Papers structures were unpicked. A verified, public beneficial ownership register would make the same connection in minutes.
Trade-related: transit fraud.
Duty-free transit fuel is being offloaded and sold on the local market; the tankers run light to the exit border to be signed off. The standard response is upfront, refundable duty paid as if for home consumption, plus electronic cargo tracking on fixed routes, accepting the trade-off that some genuine transit may move to a rival corridor.
This is the thread the tool has followed. Each step led to the next: what an IFF is, the types it takes, the sectors it runs through, why it persists, how it is found, and what stops it. Tap each to bring it back.
Tap any row to open or close it.
Figures and cases: Global Financial Integrity, illicit-flow estimates for developing countries; UNCTAD, Economic Development in Africa Report 2020; the Panama Papers (ICIJ, 2016); SwissAid analysis of UN trade data (2024); public reporting on the Marange diamond fields (2016) and on revenue-authority drone operations and fuel-transit measures. IFF estimates are risk indicators, not audited counts.